    [Nonpatent document 1] DENSO Technical Review VOL. 6 NO. 1 2001 P 29-34
In Japan, there is generally used a navigation apparatus-use map data format, which is called the KIWI/A format (for example, refer to Nonpatent document 1). In this format, a road on a map is represented in units of links; each link has reference information of a starting node and a terminating node. In addition, the map data based on this format is stratified hierarchically into several layers (i.e., levels) to improve a speed in route retrieval. The lowest layer is provided with links, each of which is assigned with a link ID for identification. In particular, for the purpose of enhancing an efficiency in drawing of roads having an identical attribute and reducing a data volume, the group of the links, which have the identical attribute and are connected serially, is stipulated as a multilink (also referred to as a link string), and links constituting the multilink are assigned with link IDs having consecutive values. Thereby, when expressing the multilink in a higher level layer, the link IDs of the starting end and terminating end of the multilink are used to identify the links within the multilink.
The map format, which is generally used such as the format of the above-mentioned KIWI/A format, is prepared by dividing into route retrieval data and map drawing data so as to provide suitable formats to applications which uses the two data, respectively. This further raises access speeds to the data and provides a system with short processing time.
In this regard, however, preparing the KIWI/A format having the two divided data groups of the route retrieval data and the map drawing data results in necessity of doubly having the identical data (for example, road information etc.), posing a disadvantage to increase the volume of the map data. In addition, the route retrieval data and map drawing data have division sizes (i.e., a parcel versus a region) different from each other for the optimization. This also poses a disadvantage to need to take time and effort in addition and deletion.
Furthermore, when a route retrieval is made using the map data based on the above format, the connection from a lower level layer to a higher level layer is made by node information (higher level connection node). For instance, suppose a route retrieval that is made from a link of a higher level like in a case where a road also existing in a higher level layer such as a notional road is selected as a starting point or a destination. In this route retrieval, a route calculation process is first made in links of a lower level layer; then, a higher level connection node is found to thereby enable the transfer from the lower level layer to the higher level layer.
However, in the route retrieval using the map data based on the above format, the reciprocal transfers between the higher level layer and the lower level layer are necessary when the route retrieval is made from a higher level layer. This needs a time in processing of the route retrieval, posing a problem. In addition, finding the higher level connection node also needs a time.